On Mon, Jun 08, 1998 at 08:43:44PM -0400, Paul Miller wrote: > > BTW they seem very suceptible to IO bandwidth.. a few things: > > 1) never burn files that are not stored on a local drive > > 2) put the writer on its own IDE controller, it should not share > > a controller with another drive that is being used > > 3) SCSI is better than IDE :) > > If the drive has a 1MB or 2MB buffer and is only writing at 2X or 4X, why > does it matter how fast the interface is? Most drives are 2X, which is > something like 300KB/s.. My motherboard supports up to 20MB/s on both of > its IDE channels. So even if the drive is on a shared channel, it'll > still be able to continously write at 300KB/s, right?
Sure, but continuous, for the whole 37 odd minutes to write a full CD? Yesterday I burnt a full CD for someone, with the files located on their PC, over Windows networking (I was running NT, they were running 95). I recorded it on the fly, without making an image first; the test phase (the first half, until we cancelled it anyway) kept the buffer 100% full. But on the actual CD, the last 100mb of files are all unreadable. Ethernet is 10mbit/s, ie 1mbyte/sec, which is >> 600kb/s for a 4X write, you say? A CDR benchmarking program says that the transfer rate off the Windows machine was under 600kb/sec on average. Hamish -- Hamish Moffatt, [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED] Latest Debian packages at ftp://ftp.rising.com.au/pub/hamish. PGP#EFA6B9D5 CCs of replies from mailing lists are welcome. http://hamish.home.ml.org -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
